How Redress Compliance delivered USD 8.5 million in savings over three years and a 35% reduction in annual AWS costs for a 15,000-employee global technology company through comprehensive usage analysis, Reserved Instance optimisation, resource rightsizing, waste elimination, competitive benchmarking, and enterprise agreement renegotiation.
Part of the AWS Contract Negotiation series
📚 AWS Negotiation Case StudiesFor all AWS contract negotiation case studies and services, start there.
The company was a global technology firm headquartered in the United States with development centres, client delivery teams, and data centres across North America, Europe, and Asia-Pacific. It operated a portfolio of SaaS products, managed cloud services for enterprise clients, and maintained a substantial internal technology platform supporting 15,000+ employees. AWS was the company’s primary cloud provider, hosting everything from production SaaS workloads to development sandboxes to data analytics pipelines.
Over the preceding three years, annual AWS spend had grown from USD 5 million to over USD 14 million — a 180% increase driven by SaaS platform expansion, new client onboarding, the growth of data and ML workloads, and the proliferation of development and testing environments across engineering teams. This growth had occurred without a corresponding cloud financial management strategy: no centralised commitment programme, no systematic rightsizing, no automated resource lifecycle management, and no enterprise-level agreement with AWS that reflected the company’s scale and growth trajectory.
The consequences were predictable. Individual business units and engineering teams provisioned resources independently, creating significant overprovisioning and waste. Compute workloads ran almost entirely on on-demand pricing, paying 40–60% premiums over committed rates for workloads that were perfectly predictable. Storage and data transfer costs grew unchecked as teams accumulated data without lifecycle policies. And the company’s existing AWS agreement provided no volume discounts, no pricing protections, and no flexibility to adapt commitments as business priorities shifted.
15,000+ employees with operations across North America, Europe, and Asia-Pacific. SaaS product portfolio, managed cloud services for enterprise clients, data analytics platforms, and internal corporate systems — all running on AWS as the primary cloud provider.
Annual AWS spend had escalated from USD 5 million to over USD 14 million driven by SaaS expansion, new client onboarding, ML/data workloads, and dev/test environment proliferation — without a corresponding cloud financial management strategy or commitment programme.
Nearly all compute ran on on-demand pricing, paying 40–60% premiums over committed rates. No Reserved Instances, no Savings Plans, no enterprise discount programme. Storage and data transfer costs grew unchecked without lifecycle policies or governance.
Individual business units and engineering teams provisioned AWS resources independently. No centralised cost management, no tagging governance, no automated lifecycle management, and no enterprise agreement reflecting the company’s USD 14 million annual scale.
AWS contract negotiations for large technology companies follow distinct patterns that differ significantly from traditional enterprise software agreements. Understanding these dynamics is essential to achieving an outcome that reflects the company’s genuine requirements and commercial leverage rather than AWS’s default pricing structure.
AWS’s pricing model defaults to on-demand rates — the highest pricing tier for any service. Companies that grow organically on AWS without implementing commitment strategies (Reserved Instances, Savings Plans, Enterprise Discount Programme) pay 40–60% more for compute and 20–30% more for storage than they would under committed pricing. Unlike traditional software licensing where overprovisioning is a procurement decision, on AWS the overspend compounds continuously as consumption grows. A company spending USD 14 million annually on-demand may be paying USD 4–6 million more than necessary.
AWS offers Enterprise Discount Programmes (EDPs) that provide percentage discounts against committed spend levels. However, EDP terms vary significantly based on the company’s negotiating position, growth trajectory, competitive alternatives, and the specific services consumed. AWS’s initial EDP proposals typically offer 5–10% discounts when 15–25% is achievable for companies with the right leverage. Without market intelligence on achievable EDP terms, companies accept AWS’s initial offer believing it represents the best available pricing.
Technology companies’ engineering-driven cultures prioritise speed and experimentation — spinning up environments rapidly, over-provisioning for peak headroom, and rarely decommissioning resources when projects conclude. At enterprise scale, this creates a persistent waste layer of 15–25% of total spend: orphaned resources, over-provisioned instances, unattached storage volumes, idle load balancers, and development environments running 24/7 when they’re used only during business hours.
We structured the engagement across six phases designed to establish complete visibility into the company’s AWS consumption, identify every optimisation opportunity, benchmark against industry peers, develop a negotiation strategy, negotiate the enterprise agreement, and implement governance to sustain the optimised position.
We conducted a comprehensive analysis of AWS usage across all accounts, regions, and business units. This included EC2 instance utilisation analysis by workload type (production SaaS, client workloads, dev/test, analytics, corporate IT), storage analysis (S3, EBS, EFS) including lifecycle assessment and access frequency, data transfer cost analysis across regions and availability zones, managed service usage (RDS, ElastiCache, Redshift, Lambda, EKS) and pricing tier assessment, and complete resource tagging audit to establish cost attribution by business unit, client, and project.
We translated the usage analysis into specific, quantified optimisation actions: rightsizing over-provisioned instances, implementing Reserved Instance and Savings Plan coverage for predictable workloads, eliminating unused resources, transitioning storage to appropriate tiers, and implementing scheduling for non-production environments.
We benchmarked the company’s AWS costs and contract terms against comparable global technology companies. This included per-unit compute costs by instance family and commitment level, EDP discount rates relative to commitment size and growth trajectory, data transfer and storage pricing benchmarks, and contract structural terms (commitment flexibility, pricing protections, credit provisions).
We developed a data-driven negotiation strategy combining the optimisation findings (demonstrating the company’s actual committed spend potential), the benchmarking data (documenting achievable EDP terms), competitive positioning (the company’s active evaluation of Azure and GCP for specific workloads), and growth projections (the company’s three-year AWS consumption trajectory as commercial leverage).
We led structured negotiations with AWS, presenting the company’s optimised consumption position and negotiating an Enterprise Discount Programme with terms that reflected the company’s scale, growth, and commercial leverage rather than AWS’s standard initial offer.
We designed and implemented a cloud financial management framework including automated monitoring, cost allocation, optimisation governance, and regular review processes to sustain the optimised position.
The usage analysis revealed a significant layer of waste across the company’s AWS estate — resources that were consuming budget without delivering operational value. Eliminating this waste was the foundation for the broader optimisation and negotiation strategy, as it demonstrated to AWS that the company’s actual committed spend would be based on genuinely required resources, not inflated consumption.
The analysis identified 340+ orphaned resources across all accounts: unattached EBS volumes from terminated instances, idle Elastic IP addresses, unused load balancers from decommissioned projects, abandoned RDS snapshots beyond retention requirements, and S3 buckets with no active access. These resources had accumulated over three years of engineering-driven provisioning without lifecycle management. Cleanup saved approximately USD 680,000 annually.
Forty-two percent of EC2 instances were provisioned at a higher instance type than their actual CPU and memory utilisation warranted. Development environments running on m5.2xlarge instances averaged 12% CPU utilisation; production instances provisioned for peak headroom averaged 25–35% utilisation outside peak periods. Rightsizing these instances to match actual workload requirements — moving to smaller instance types or graviton-based alternatives — saved approximately USD 620,000 annually.
Development, testing, staging, and QA environments ran 24/7 despite being used only during business hours (approximately 50–60 hours per week versus 168 hours). Implementing automated start/stop scheduling for non-production environments reduced their compute costs by approximately 65%. Annual savings: USD 480,000.
Approximately 180 TB of S3 data was stored in Standard tier despite having not been accessed in over 90 days. Implementing S3 Intelligent-Tiering for actively managed data and lifecycle policies transitioning infrequently accessed data to S3 Glacier Instant Retrieval or S3 Glacier saved approximately USD 320,000 annually. Additionally, EBS volume optimisation (gp3 migration from gp2, right-sizing provisioned IOPS) contributed to the total savings.
With waste eliminated and resources rightsized, we designed a commitment strategy that maximised savings on the company’s genuinely required workloads. The strategy combined Reserved Instances for stable production workloads, Savings Plans for variable compute, and strategic placement decisions to optimise data transfer costs.
Before optimisation: Nearly all compute ran on on-demand pricing. No Reserved Instances, no Savings Plans, no commitment programme. The company was paying the highest available rate for every hour of compute consumed.
Our strategy: We categorised every workload by stability and predictability, then applied the optimal commitment mechanism: 3-year Reserved Instances for production SaaS workloads (stable, 24/7), 1-year Savings Plans for client-facing workloads (predictable but growing), and on-demand for genuinely variable workloads (development spikes, batch processing, experimental infrastructure).
Beyond compute and storage, the company’s managed service consumption offered significant optimisation opportunities. Engineering teams had adopted managed services independently, creating pricing tier mismatches and service overlaps that accumulated substantial unnecessary cost.
The company operated three Redshift clusters for analytics workloads. Two were provisioned at ra3.4xlarge with consistently low utilisation, while the third was appropriately sized. We consolidated the underutilised clusters into a single right-sized cluster with Redshift Serverless for variable analytical workloads, reducing annual Redshift costs by approximately USD 220,000.
Fourteen RDS instances were provisioned for peak capacity that occurred during quarterly reporting and end-of-month processing. We migrated non-critical databases to RDS Aurora Serverless v2, which scales automatically based on demand, and right-sized the remaining production instances. Annual savings: approximately USD 180,000.
Lambda functions were over-provisioned with memory allocations 2–4x above actual requirements, and EKS clusters ran with substantial unused node capacity. We right-sized Lambda memory allocations using AWS Compute Optimizer data and implemented Karpenter for EKS cluster autoscaling, reducing costs by approximately USD 140,000 annually.
Armed with the optimised consumption baseline, competitive benchmarking data, and growth projections, we entered structured negotiations with AWS. The negotiation produced an Enterprise Discount Programme with terms that reflected the company’s genuine commercial leverage rather than AWS’s standard initial offer.
AWS’s initial EDP proposal offered a 7% discount against a USD 10 million annual commitment. Our counter-position, backed by independently verified consumption data and market benchmarks, demonstrated that the company’s optimised baseline justified a significantly higher discount rate and more flexible commitment terms.
By eliminating USD 2.1 million in waste before negotiating, we demonstrated to AWS that the company’s committed spend would be based on genuinely required resources. This credibility prevented AWS from inflating commitment targets to absorb the waste we had identified. The optimised baseline also showed AWS that the company understood its own consumption at a granular level — signalling that inflated proposals would be challenged with data.
The company was actively evaluating Azure for specific Microsoft-integrated workloads and GCP for ML/data analytics pipelines. These were genuine evaluations, not negotiating bluffs. We presented the company’s multi-cloud strategy as factual context: AWS would retain the majority of workloads, but competitive pricing was essential to prevent workload migration to alternative providers. This positioned competitive EDP terms as a retention investment for AWS.
The company’s three-year growth projections showed AWS consumption growing to USD 18–22 million annually through SaaS expansion, new client onboarding, and data platform growth. We framed the EDP negotiation around this growth trajectory: AWS was securing a three-year relationship with a technology company whose consumption would grow 50–80%. Preferential EDP terms were the mechanism to ensure this growth materialised on AWS rather than migrating to competitors.
| Category | Annual Savings | Three-Year Impact |
|---|---|---|
| Waste elimination (orphaned, rightsizing, scheduling, storage) | USD 2.1M | USD 6.3M |
| Commitment strategy (RIs, Savings Plans, data transfer) | USD 3.2M | USD 9.6M |
| Managed service rationalisation | USD 540K | USD 1.62M |
| EDP negotiated discounts (incremental) | USD 780K | USD 2.34M |
| Total (net of overlap with commitment) | ~USD 4.9M annual | USD 8.5M over 3 years |
“Redress Compliance’s expertise in AWS negotiations delivered exceptional value. Their strategic approach helped us achieve significant cost savings while aligning our contract with our operational and innovation goals. Their support was critical in optimising our cloud strategy.” — CTO, Global Technology Company
The negotiation established the optimal commercial framework. Sustaining the savings required a cloud financial management discipline that the company had lacked. We implemented a governance framework designed for the realities of a 15,000-employee technology company where engineering teams provision resources autonomously and consumption patterns change continuously.
We implemented AWS Cost Explorer and Cost Anomaly Detection with business-unit-level budgets and automated alerting. Each business unit received monthly cost reports with trend analysis, anomaly flags, and optimisation recommendations. Executive dashboards provided real-time visibility into total AWS spend, commitment coverage, and savings plan utilisation — ensuring that leadership maintained awareness of the company’s cloud investment efficiency.
We enforced mandatory resource tagging through AWS Service Control Policies and Config Rules. Every resource required business-unit, project, environment (production/staging/dev/test), and owner tags. Resources without compliant tags were flagged within 24 hours and escalated for remediation. This tagging discipline enabled precise cost attribution and made waste visible to the teams creating it.
We established quarterly cloud optimisation reviews that assessed RI and Savings Plan coverage and utilisation, instance rightsizing opportunities using Compute Optimizer data, storage lifecycle policy effectiveness, orphaned resource accumulation, and managed service pricing tier alignment. Each review produced specific, quantified actions with assigned owners and completion deadlines.
We delivered cloud financial management training for engineering leads, product managers, and finance stakeholders covering AWS pricing models, commitment strategies, resource lifecycle management, cost-aware architecture decisions, and the quarterly review process. The training established a FinOps culture where cost efficiency was treated as a shared engineering responsibility, not just a procurement concern.
This engagement reinforced patterns we observe consistently in technology company AWS contract negotiations. The specific savings scale with consumption, but the underlying dynamics — on-demand overspend, resource sprawl, absence of commitment strategy, and AWS’s conservative initial EDP offers — appear in virtually every technology company we advise.
Eliminating USD 2.1 million in waste before negotiating prevented AWS from anchoring the EDP commitment to an inflated consumption baseline. Companies that negotiate first and optimise later commit to spend levels that include waste — paying for resources they don’t need at a discounted rate rather than eliminating the waste entirely. Optimise first, then negotiate commitments against the genuine consumption baseline.
This company was paying on-demand rates for USD 14 million in annual consumption when USD 9+ million qualified for committed pricing at 25–45% discounts. The absence of a commitment strategy was costing the company approximately USD 3.2 million annually. Every technology company spending more than USD 1 million annually on AWS should implement a structured commitment programme — the savings are immediate and substantial.
AWS’s initial 7% EDP offer was well below the achievable rate for a company with this consumption level and growth trajectory. Without market intelligence on comparable EDP terms, companies accept the initial offer believing it represents competitive pricing. Independent benchmarking and negotiation expertise typically achieves 15–25% EDP discounts for companies at this scale.
Fifteen to twenty-five percent of this company’s AWS spend was waste. Without automated governance (tagging enforcement, lifecycle policies, scheduling, anomaly detection), waste regenerates continuously as engineering teams provision new resources. Cloud financial management is not a one-time exercise; it requires permanent governance infrastructure.
The company’s genuine evaluation of Azure and GCP for specific workloads provided credible competitive leverage. AWS understands that technology companies can and do operate across multiple cloud providers. Competitive positioning that is grounded in real evaluations — not empty threats — is the most effective lever for achieving preferential EDP terms.
The advisory investment represented less than 3% of the three-year savings achieved. Without independent cloud financial management and negotiation expertise, the company would have continued paying on-demand premiums, accumulating resource waste, and accepting AWS’s conservative EDP terms. The information asymmetry between AWS’s account team and a company without specialised cloud commercial expertise is substantial.
AWS contract negotiations for technology companies involve millions of dollars in annual cloud spend, complex consumption patterns across dozens of services, and commercial dynamics that differ fundamentally from traditional enterprise software agreements. AWS’s account teams have deep visibility into the company’s consumption data and are incentivised to maximise committed spend. Without independent advisory, companies negotiate with less information about their own optimisation potential than AWS has about their consumption patterns.
In this engagement, the company’s internal teams — engineering, finance, and procurement — were focused on building products and serving clients. They lacked the specialised expertise to conduct service-level usage analysis across 50+ AWS services, design optimal commitment strategies balancing RI coverage with flexibility, benchmark EDP terms against comparable technology companies, quantify the specific savings achievable through rightsizing, scheduling, and lifecycle management, and negotiate with AWS from independently verified data. The difference was USD 8.5 million over three years.
Redress Compliance’s team includes specialists in AWS pricing models, commitment strategies, managed service optimisation, and EDP negotiation. This expertise enables the depth of consumption analysis and commercial strategy that transforms contract outcomes for technology companies at enterprise scale.
We maintain current benchmarking data across technology companies’ AWS EDP terms, discount rates, and contract structures. This intelligence ensures our clients negotiate with full awareness of achievable terms — not the conservative initial offers that AWS presents as competitive pricing.
Redress Compliance has no commercial relationship with AWS — no partner status, no resale revenue, no referral commissions. Our consumption analysis, optimisation recommendations, and negotiation strategy are exclusively aligned with our clients’ interests. This independence is essential when recommending multi-cloud strategies or challenging AWS’s proposed commitment levels.
“Technology companies are typically overpaying for AWS by 25–40%. The combination of on-demand pricing defaults, resource sprawl, absent commitment strategies, and conservative EDP offers means that companies negotiate contracts based on inflated consumption and uninformed commercial expectations. Independent usage analysis, optimisation, benchmarking, and negotiation reverses this dynamic entirely.”
In our experience with technology company AWS engagements, companies typically achieve 25–40% reductions in annual AWS costs through a combination of waste elimination (15–25% of spend), commitment strategy implementation (25–45% savings on committed workloads), managed service rationalisation, and EDP negotiation. This company achieved a 35% reduction and USD 8.5 million in savings over three years — consistent with the range we see for technology companies at this scale.
An AWS EDP is a contractual agreement where a company commits to a minimum annual AWS spend level in exchange for a percentage discount applied across all AWS services. EDP terms vary based on commitment size, growth trajectory, negotiating leverage, and the specific services consumed. AWS’s initial EDP offers typically range from 5–10%, but 15–25% discounts are achievable for companies with the right leverage and negotiation approach. EDPs are typically structured as 1–3 year agreements with annual commitment escalators.
Always optimise before negotiating. If you negotiate an EDP based on your current inflated consumption (including waste and on-demand pricing), you commit to spend levels that include resources you don’t need. Optimising first establishes your genuine consumption baseline, prevents AWS from anchoring commitments to inflated spend, and demonstrates to AWS that you understand your own consumption at a granular level — which strengthens your negotiating position significantly.
Reserved Instances provide the deepest discounts (up to 45% for 3-year commitments) but are tied to specific instance types, regions, and operating systems. Savings Plans provide slightly lower discounts (up to 30%) but offer flexibility to apply across instance families, regions, and compute services (EC2, Fargate, Lambda). The optimal strategy combines both: RIs for stable, predictable workloads where the instance type is unlikely to change, and Savings Plans for workloads that may evolve in terms of instance type or region.
Sustained optimisation requires permanent governance infrastructure: mandatory resource tagging enforced through Service Control Policies, automated scheduling for non-production environments, lifecycle policies for storage, anomaly detection and alerting, and quarterly optimisation reviews. Without these mechanisms, waste regenerates within 3–6 months as engineering teams provision new resources. The governance framework is as important as the initial optimisation.
Typically 10–14 weeks depending on the complexity of the AWS estate and the number of accounts and business units involved. This engagement completed in 12 weeks: usage analysis (3 weeks), cost optimisation (2 weeks), benchmarking (1 week), negotiation strategy (1 week), contract negotiation (3 weeks), and governance implementation (2 weeks). We recommend starting 4–6 months before any existing EDP or commitment expiration to maximise optimisation implementation before the new agreement begins.
No. Redress Compliance is a 100% independent advisory firm with no commercial relationship with AWS or any other cloud provider. We do not resell AWS services, hold AWS partner status, or earn referral commissions. This independence ensures our consumption analysis, optimisation recommendations, and negotiation strategy are exclusively aligned with our clients’ interests — including recommendations to evaluate alternative cloud providers where appropriate.
Let Redress Compliance run a comprehensive AWS usage analysis and deliver a negotiation strategy that protects your interests and maximises savings.
This case study is part of our AWS Negotiation Case Studies series. Explore related case studies and services: